INDEX

Explanations

phrases related to negative occurrences or disasters

oai_token-act-pair · gpt-3.5-turbo

phrases indicating events or situations that have occurred over time, especially those with significant impact

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.hook_resid_pre

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

24,576

Data Type

torch.float32

Hook Point

blocks.10.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

eele

-0.76

erence

-0.74

 guessing

-0.67

 Charisma

-0.65

OTAL

-0.65

ISSION

-0.65

 specificity

-0.64

 representations

-0.64

Reviewer

-0.63

 nods

-0.63

POSITIVE LOGITS

 ravaged

1.51

 rocked

1.46

 erupted

1.43

 engulfed

1.43

 raged

1.39

 engulf

1.39

 devastated

1.37

 swept

1.35

 ensued

1.35

 gripped

1.35

Activations Density 0.237%

phrases related to negative occurrences or disasters

phrases indicating events or situations that have occurred over time, especially those with significant impact

No Comments

No Known Activations