INDEX

Explanations

words related to slight adjustments or changes

oai_token-act-pair · gpt-3.5-turbo

incremental changes or modifications

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.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

nor

-0.78

dest

-0.73

all

-0.72

ahu

-0.71

anca

-0.68

ios

-0.67

mining

-0.67

ondon

-0.66

 Nations

-0.65

tics

-0.65

POSITIVE LOGITS

 slightly

3.16

 marginally

2.27

 somewhat

2.05

 mildly

1.79

 considerably

1.78

lightly

1.77

 moderately

1.74

 slight

1.74

 Slightly

1.74

 faintly

1.65

Activations Density 0.017%

words related to slight adjustments or changes

incremental changes or modifications

No Comments

No Known Activations