INDEX

Explanations

phrases related to surprise or realization

oai_token-act-pair · gpt-3.5-turbo

the phrase "Was" in the context of narratives or reflections

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

ogle

-0.66

aim

-0.61

 afford

-0.59

 disruptive

-0.58

 come

-0.58

 passively

-0.58

ICA

-0.57

acci

-0.57

 taking

-0.56

 investing

-0.56

POSITIVE LOGITS

Was

3.42

Was

2.91

 Were

1.93

Did

1.88

was

1.73

 Didn

1.65

WAS

1.64

Had

1.60

Is

1.59

 Would

1.57

Activations Density 0.009%

phrases related to surprise or realization

the phrase "Was" in the context of narratives or reflections

No Comments

No Known Activations