INDEX

Explanations

text related to patterns or recurring sequences.

oai_token-act-pair · gpt-3.5-turbo

various types of patterns across different contexts

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

 affair

-0.89

 announcement

-0.87

 requirement

-0.86

 complaint

-0.86

 lawsuit

-0.85

 exception

-0.82

 issue

-0.82

ittee

-0.82

 invitation

-0.81

 embargo

-0.81

POSITIVE LOGITS

 patterns

2.11

 Patterns

1.64

 concepts

1.63

 scenarios

1.62

 interpretations

1.61

 perspectives

1.61

 behaviors

1.56

 modes

1.55

 explanations

1.55

 meanings

1.55

Activations Density 0.088%

text related to patterns or recurring sequences.

various types of patterns across different contexts

No Comments

No Known Activations