INDEX

Explanations

instances where attention is drawn towards specific details or issues

oai_token-act-pair · gpt-3.5-turbo

topics related to neglect or lack of attention given to important issues

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 7-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.7.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.7.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

sts

-0.68

fu

-0.66

gran

-0.65

Rating

-0.64

anian

-0.61

imen

-0.60

rain

-0.60

 Royale

-0.59

zer

-0.59

tti

-0.59

POSITIVE LOGITS

 adequately

0.84

nor

0.78

 nuance

0.76

 whatsoever

0.75

 mention

0.74

 altogether

0.73

 bothered

0.73

 nuances

0.73

 ãĤµãĥ¼ãĥĨãĤ£ãĥ¯ãĥ³

0.73

 cues

0.70

Activations Density 0.491%

instances where attention is drawn towards specific details or issues

topics related to neglect or lack of attention given to important issues

No Comments

No Known Activations