INDEX

Explanations

phrases beginning with "every time" where a negative event or action is described

oai_token-act-pair · gpt-3.5-turbo

occurrences of the phrase "every time" in various contexts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.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

riad

-0.81

otto

-0.79

anse

-0.77

querque

-0.77

yright

-0.75

MSN

-0.73

ourt

-0.72

emetery

-0.72

XT

-0.72

ordes

-0.69

POSITIVE LOGITS

 consuming

0.83

frames

0.79

soever

0.77

period

0.77

 lapse

0.73

frame

0.73

 frame

0.72

bomb

0.67

 someone

0.66

 fluor

0.66

Activations Density 0.027%

phrases beginning with "every time" where a negative event or action is described

occurrences of the phrase "every time" in various contexts

No Comments

No Known Activations